From: AC
To: Editor

 

While browsing the web I came across your site and in particular the comments regarding Simon Williams and his Associative Database system "Sentences".

 

You may be interested to know that the concepts Williams now claims as his own were published on the web by me around eight years ago on my web site first at Cityscape.co.uk (now defunct) then at Pipex Dial, dspace.dial.pipex.com, (when that acquired Cityscape). I started to jot some ideas down about nine years ago as the basis of a book; see the attached document. At the time I was interested in AI and thought that the AI attempts of the day were flawed in the way they handled data. I thought that in order to approach a human capability there needed more thought on the way data was stored, and that the key to AI was not information processing, but information storage. You may also be interested that I was an ORACLE Development Team Manager at the time (as well as doing an SSADM qualification) so was reasonably informed about data structures and the relational data model.

 

If an information storage system could mimic the way we stored information then that information would become processable in similar ways to the ways in which we think. Thus I felt that processes like leaps of intuition, creative and lateral thinking are consequences of the way in which we store information. To store data in conventional RDBMS structures fundamentally makes such processes difficult and so AI needed to develop new storage structures.

 

I therefore proposed Associative storage as a storage model, which made possible AI type information processing. I got as far as designing two basic necessary structures; a data or atom structure; which stored all the atoms of information and a link table, which stored the links between these atoms. I wasn't sure whether the associative links should have attributes; e.g. the type of association they were or whether to store this again by association. At the time I had difficulty getting hold of any database programming add-ons to try out my ideas, though eventually I got the Borland Paradox Engine in March 1994.

Since I didn't develop these ideas further I am happy to wish Williams well in his use of them, but I am not happy for him to claim that he invented them. After reading your comments I looked at his website and he says he is claiming US patent rights on these ideas. It is possible he did independently come up with similar ideas to me, but I believe I was first (in 1992/1993) and if anyone has an archive of PIPEX from 1994 they will see this. I also have a notebook in which I jotted this all down, which should be forensically dateable too.

 

As part of my dabblings around this I started toying with the idea of creating a company called Alogic (short for Associative Logic). I have use the email address achapman.alogic@virgin.net for around four years and my website on Cityscape and Pipex Dial was headed with an ALogic logo. I also hosted a Local Government web directory called the "Local Gov One stop Shop" run under this Alogic banner.

 

Therefore I think I can prove my case on this. Up until now I haven't been too bothered, since my deliberations on this concluded that there were a series of problems with taking it further; one of which was the problems in representing the data physically in a way that was more efficient than conventional data structures. I do object to Williams’s arrogance, though, as if he has come up with something that is totally original, without crediting his sources. I would also object to being prevented from talking about something I thought up because he has the copyright on my ideas; whether or not they are flawed.

 

From: Fabian Pascal
To: AC

 

I am skeptical about your approach for various reasons.

 

Given the nonsense that Williams is throwing around, I very much doubt that he is capable of coming up with anything of value on his own. This was reinforced by the fact that when I asked him for a definition of a data model, he sent me something that he lifted as is from other sources.

 

If I were you, I would make public what you just told me -- if true, you should not let him get away with it. I can post this exchange on the site if that's OK with you.
 
 

From: AC
 

I am quite happy for you to post what I sent you the other day. I posted my original thoughts in a spirit of open science, for others to use and build on, if they wanted to. I expect the usual quid pro quo, that if people build on your ideas, they at least have the courtesy to acknowledge they are doing so. I was prepared to give Williams the benefit of the doubt, that he independently thought this all up. But the more I learn of him the less likely this seems.

 

I appreciate your skepticism, which is another reason why I didn't feel like confronting Williams. As I said in my previous email I felt there were holes in the 'Associative Model' (if I can be so grandiose), which meant I didn't particularly want to be identified with it, and certainly didn't want to be in the position of defending it against experts.
 
 

From: Fabian Pascal

Those who cannot come up with ideas of their own tend to offer no such courtesy; those who are self-assured do.

Whatever its pros and cons, your concept does not seem to be a 'data model' in the sense in which the R model is; there is a lot of confusion as to models out there, see Date's Models, Models Everywhere, Nor Any Time to Think. Williams does not seem to comprehend that even when he is using others' definition for what a data model is. That indicates a problem more serious than just lack of knowledge.

 

 

Posted 08/04/02

 

 

 

[ABOUT] [QUOTES] [LINKS]